LG AI Research and CSE build on successful partnership

A delegation from LG AI Research spent a day on the University of Michigan North Campus on Oct 11, 2023 to review LG-funded research projects in process, meet with leadership and faculty to discuss future collaborations, and interact with students working in the AI space about opportunities at LG AI Research.

LG AI Research, with global headquarters in Seoul, South Korea, has taken a number of steps to develop ties with the University of Michigan to establish a leadership position in North America for the responsible deployment of AI and in pursuit of its mission to advance AI such as Deep Reinforcement Learning, 3D Scene Understanding, and Reasoning with a Large-scale Language Model and Bias & Fairness related to AI ethics, which LG AI Research sees as the basis for creating AI that thinks and judges on its own.

This has included the establishment of the North American LG AI Research Center in Ann Arbor. Officially opened at a ceremony in March of 2022, the center is located in close proximity to the U-M North Campus where CSE is housed. Prof. Honglak Lee, a faculty member in CSE, also serves as Chief Scientist of AI at LG AI Research and Head of LG AI  Research Ann Arbor Center.

Dr. Kyunghoon Bae, Chief of LG AI Research and Senior VP at LG, led the recent campus visit; Prof. Lee was also in attendance. The two met with Associate Dean for Research and Louise Ganiard Johnson Professor of Engineering Eric Michielssen, and with Richard H. Orenstein Division Chair of Computer Science and Engineering and Lynn A. Conway Collegiate Professor of Computer Science and Engineering Michael Wellman to confer about the LG AI Research – University of Michigan relationship.

Additional faculty who met with the LG AI Research visitors included:

• Prof. Rada Mihalcea and Prof. David Jurgens, who reviewed their work on two LG AI funded projects entitled “Knowledge Grounded Personalized Dialogue Generation” and “Towards People-Centric Language Models with Self Alignment and Prompt Refinement.”
• Prof. Joyce Chai, who reviewed her work on two projects entitled “Multimodal Language Models for Coherent Physical Commonsense Reasoning” and “Multimodal Large Language Models for Situated Theory-of-Mind Reasoning in Embodied Agents.”
• Prof. Justin Johnson, who reviewed his work on three projects entitled “Visual Representation Learning from Language and 3D,” “Self-Supervised Concept Discovery in Uncurated Multimodal Data,” and “Efficient and Scalable Text-to-3D Generation.”
• Prof. Lu Wang, who reviewed her work on three projects entitled, “Reasoning with Large Models,” “Knowledge Grounded Scientific Reasoning,” and “Effective and Fine-Grained Feedback for Enhanced Language Model Reasoning and Alignment.”
• Prof. Qiaozhu Mei, who reviewed his work on a project entitled, “Teach Large Language Models to Innovate.”

Reflecting on the day’s activities, Prof. Lee said, “We are excited by the progress of the research currently underway and the future opportunities that exist to expand on this partnership. LG AI Research and the University of Michigan share a vision for advancing interdisciplinary AI research that will have a positive impact on the world.”

Associate Dean Michielssen also commented that, “Through this partnership with LG AI Research, our students and faculty are able to trade expertise with our counterparts at one of the most innovative AI research centers in the world. It is exciting to see this relationship blossom.”